Comments on the paper 'Using G2 to measure inequality in Brazil' by Tina Faulkner

摘要

The results presented in the article are consistent with other reported measurements of the recent dynamics of income inequality in Brazil, which point to a sustained decline starting in the 2000s. In addition to the Gini index of income, the authors make use of another inequality measurement - the G2, in which the Gini index is multiplied by the mean/median ratio, to detect changes that favor the upper income groups -, that corroborates the trend. To the extent that the income share of the richest 10 percent has been shrinking while that of the poorest has been increasing, the G2 does not reveal any undesirable trend and is highly correlated with the Gini. Another point mentioned in the article is the poor quality of the data that captures the income share of the poorest 10 percent in household surveys, which suggests systematic under reporting of income for this group. This may lead to overestimation of poverty and inequality. Some people then suggest that these should be calculated on the basis of consumption rather than income. The authors then bring in existing consumption-based inequality estimates for one year and find them to be lower than the income estimate. But the absence of a time series prevents them from identifying a similar trend to that of income inequality.